Copy of Copy of AI Application in HSC Belfast Trust

JT
by John Totten0 views·Updated Apr 3, 2022
Create your own

Chapters

Audio TranscriptAuto-generated

  1. Yeah.
  2. Here we will be discussing the generated synthetic data solution.
  3. Hm.
  4. Here it will be discussing the generated synthetic data solution.
  5. What a I generated synthetic data.
  6. Well,
  7. it's artificially manufactured data which replicate characteristics of already
  8. existing data without revealing the original data set.
  9. Original data is replaced, not disguised or modified.
  10. How is it created?
  11. While there are many ways our generated, pathetic data can be created,
  12. however,
  13. the most common involved algorithms learning the structure
  14. of the original data through training data sets.
  15. Once learned, the model can generate samples of synthetic data.
  16. The most prominent AI approach in this field
  17. is a generative adversarial networks,
  18. most commonly known as Gans.
  19. We can see a quick illustration of how those work,
  20. Uh,
  21. what are some of the benefits of using synthetic data?
  22. Well, it allows us to do identify original patient data.
  23. Therefore, removing risk of sensitive data breaches which,
  24. as we mentioned before, can be a huge issue within healthcare.
  25. It also offers safe legal and ethical open data sharing for collaboration
  26. and greater reproduce ability and transparency in medical research.
  27. It also helps improve inaccuracy for machine learning models.
  28. I'm very importantly,
  29. AIDS research into rare disease is insufficient. Data is lacking in the data sets.
  30. All of the above can lead to great strides in a lead healthcare.
  31. Now let's look at a few examples of where
  32. synthetic data has been applied in healthcare before.
  33. Creation of synthetic medical imagery
  34. can be used in regulatory stress testing before deployment.
  35. An illustration as we can see here on the right of the real images of some
  36. medical issues.
  37. And on the left, where the synthetic data has allowed medical research,
  38. is to create synthetic images replicating those
  39. the N H s u synthetic data to support a
  40. I that fight covid-nineteen and other cardio vascular diseases,
  41. as well as to train ai, which diagnosed diabetic symptoms in the feet.
  42. Here we can see. The director of the N.
  43. H s has suggested that creating synthetic data set is a novel way
  44. to help train machine learning algorithms on a rich and diverse set of data
  45. whilst maintaining safety, protecting privacy,
  46. something again very important within the healthcare context.
  47. Hm.
  48. I will now be handing over to Joe, who will be discussing
  49. the
  50. uh
  51. yeah,
  52. I will now be handing over to Joe who will
  53. be discussing the responsible use of a I in healthcare.
  54. Mhm.
  55. Yeah.
  56. I will now be handing over to joke who will
  57. be discussing the responsible use of AI within healthcare.